[Bug] Message loss on bookies and brokers restart

[szkoludasebastian]

Search before asking

• [X] I searched in the issues and found nothing similar.

Read release policy

• [X] I understand that unsupported versions don't get bug fixes. I will attempt to reproduce the issue on a supported version of Pulsar client and Pulsar broker.

Version

Client version: 3.2.2, Server version: 3.2.2 On previous version also notice same behaviour, e.g. 3.1.0, 3.1.2

Minimal reproduce step

Noticed messages loss when we restart all bookies and brokers during processing the data, so data is send to some topic and then our application consumes messages from topic saves msg payload somewhere and acknowledges messages. To be more precise here are steps:

1. Simulate data streaming - send 1000000 msgs to topic A in some time, e.g. 1000000 msgs in 2 minutes.
2. During the time when messages are sent restart all bookies and brokers at once (in our case we have 8 bookies and 6 brokers).
3. In the same time application is consuming messages from topic A, saves message payload somewhere and after successful save acknowledges messages.

What did you expect to see?

No message loss

What did you see instead?

Some messages are lost. So when we send 1000000 messages, in directory where we store messages we see less than 1000000. We can't specify here how much less, because it is a very random situation. Sometimes we have all the messages, but sometimes something is missing.

Anything else?

No response

Are you willing to submit a PR?

• [ ] I'm willing to submit a PR!

[dao-jun]

It looks a little strange, did you handle the send message result?

I mean, what do you do when send message failed?

It looks you don't handle the send message result, if there are some messages failed to send, you should handle the exception and retry again

In general, Pulsar do not retry to send message automatically, it's user's responsibility

[szkoludasebastian]

I think it's not the case, because we are retrying. We handle message result and retry message if it was not sent correctly. During this test which I described we are sending messages till we won't have 1000000 correctly sent.

So we are sure that on input we have 1000000 messages.

[dao-jun]

Did you mean you send 1000000 messages, the results are succeed, but after restart broker/bookie, some of them lost?

[szkoludasebastian]

I mean we send 1000000 messages and during that time when our script was sending messages we restarted all bookies and brokers and then we were able to observe that some of messages lost. It was random situation for us, so we had to perform this test few times to observe that message loss.

[dao-jun]

can I see your script?

[szkoludasebastian]

I simplified our infrastructure a bit just to show our problem. To be more accurate our test script sends data to the application using websocket and our application sends messages to the topic. Our application responds with either ok or error results and in the case of error the script resends the message. In the case of the script and our application we see that as many messages as we expected were sent correctly.

[szkoludasebastian]

We performed many failover tests using that script and we haven't observed any message loss. Only in this case when we restarted bookies and brokers in same time we observed something like that.

[lhotari]

In the case of the script and our application we see that as many messages as we expected were sent correctly.

Which Pulsar client do you use? When you say "sent correctly", how do you define that?

[szkoludasebastian]

We use pulsar client version: 3.2.2 (org.apache.pulsar:pulsar-client-api:3.2.2).

When I say that message was sent correctly I mean that we use something like this:

public CompletableFuture<Response> sendMessage(Request request) {
	    return producer.newMessage()
                            .value(request.getPayload())	    
			    .sendAsync()
			    .thenApply(messageId -> ok(request))
			    .exceptionallyCompose(ex -> createErrorResponse(request, ex));
    }

so we are creating message based on websocket request, then we use sendAsync() and we use thenApply() to return ok response. When any exception appears then we use exceptionallyCompose() and we return error response.

[szkoludasebastian]

Here is our configuration for pulsar client (in comments I put values):

public PulsarClient createPulsarClient(ClientBuilder pulsarClientBuilder, PulsarClientProperties properties) throws PulsarClientException {
        return pulsarClientBuilder
                .serviceUrl(properties.getProxyUrl()) //pulsar://localhost:6650
                .ioThreads(properties.getHandlingBrokersConnectionThreads()) //10
                .listenerThreads(properties.getListenerThreads()) //1
                .connectionsPerBroker(properties.getMaxConnectionsPerBroker()) //10
                .connectionTimeout(parseDurationToSeconds(properties.getConnectionTimeout()), TimeUnit.SECONDS) //PT30S
                .operationTimeout(parseDurationToSeconds(properties.getOperationTimeout()), TimeUnit.SECONDS) //PT30S
                .statsInterval(parseDurationToSeconds(properties.getStatsInterval()), TimeUnit.SECONDS)
                .build();
    }

and also configuration for producer:

public Producer<byte[]> createProducer(PulsarClient pulsarClient, PulsarProducerProperties properties) {

	    ProducerBuilder<byte[]> builder = pulsarClient.newProducer(Schema.BYTES);
	    return builder.topic(properties.getTopicPublish())
                .batchingMaxPublishDelay(properties.getBatchingMaxPublishDelayMicros(), TimeUnit.MICROSECONDS) //100000
                .batcherBuilder(BatcherBuilder.KEY_BASED)
                .hashingScheme(HashingScheme.Murmur3_32Hash)
                .blockIfQueueFull(true)
                .sendTimeout(parseDurationToSeconds(properties.getSendTimeout()), TimeUnit.SECONDS) //PT30S
                .create();
 }

[lhotari]

so we are creating message based on websocket request, then we use sendAsync() and we use thenApply() to return ok response. When any exception appears then we use exceptionallyCompose() and we return error response.

@szkoludasebastian This looks correct.

[lhotari]

Some messages are lost. So when we send 1000000 messages, in directory where we store messages we see less than 1000000. We can't specify here how much less, because it is a very random situation. Sometimes we have all the messages, but sometimes something is missing.

Another point of view is to say that the messages aren't delivered to the consumer in your test scenario. It's about the same as message loss from your application perspective, but there's a subtle difference. Would you be able to check if the messages are stored ok and would be available for delivery on another subscription and consumer that is started after this failure scenario. It's possible that the message loss happens in delivery on the consumer side.

What subscription type are you using? How do you handle acknowledgements?

When the problem happens, please share the internal stats for the topic and the subscription (use pulsar-admin topics stats-internal [topicname] to get it) immediately after the problem has happened. That would reveal useful information about the internal states. Please also provide topic stats ( pulsar-admin topics stats [topicname]), just to be sure that all available information is available for diagnostics.

[szkoludasebastian]

1. Here is consumer configuration:

public Consumer<byte[]> byteConsumer(ConsumerConfigurationProperties consumerConfigurationProperties) throws PulsarClientException {
        return PulsarClient.builder()
                .serviceUrl(consumerConfigurationProperties.getProxyUrl())
                .statsInterval(consumerConfigurationProperties.getStatsInterval(), TimeUnit.SECONDS)
                .build()
                .newConsumer()
                .subscriptionMode(SubscriptionMode.Durable)
                .topic(consumerConfigurationProperties.getTopicSubscribeDecorated())
                .subscriptionName(consumerConfigurationProperties.getSubscriptionName())
                .consumerName(String.format("%s-%s", consumerConfigurationProperties.getConsumerName(), UUID.randomUUID()))
                .subscriptionType(SubscriptionType.Key_Shared)
                .keySharedPolicy(KeySharedPolicy.autoSplitHashRange())
                .subscribe();
    }

So we are using Key_Shared subscription.

2. We are handling acknowledgments this way:

public void acknowledgeMessages(List<DomainMessageId> messages) {
        List<MessageId> messageIds = messages.stream()
                .map(this::toMessageId)
                .toList();

        consumer.acknowledgeAsync(messageIds)
                .exceptionally(throwable -> {
                    log.error("ACKNOWLEDGE ERROR, messages Ids: {}", messageIds, throwable);
                    throw new IllegalStateException(throwable);
                });
    }

public void negativeAcknowledge(List<DomainMessageId> messages) {
        messages.stream()
                .map(this::toMessageId)
                .forEach(consumer::negativeAcknowledge);
    }

So we have these two methods. One for acknowledge messages and second one for negative ack. We consume messages from topic and store them in batches. When it's time to flush batch and it is correctly stored in directory which we want then we are using this method to acknowledge messages by id. When there is some error with storing data in desired directory then we negatively acknowledge messages.

3. Here we have topics partitioned-stats:

{
  "msgRateIn" : 0.0,
  "msgThroughputIn" : 0.0,
  "msgRateOut" : 0.0,
  "msgThroughputOut" : 0.0,
  "bytesInCounter" : 0,
  "msgInCounter" : 0,
  "bytesOutCounter" : 8776881,
  "msgOutCounter" : 13359,
  "averageMsgSize" : 0.0,
  "msgChunkPublished" : false,
  "storageSize" : 262236942,
  "backlogSize" : 0,
  "publishRateLimitedTimes" : 0,
  "earliestMsgPublishTimeInBacklogs" : 0,
  "offloadedStorageSize" : 0,
  "lastOffloadLedgerId" : 0,
  "lastOffloadSuccessTimeStamp" : 0,
  "lastOffloadFailureTimeStamp" : 0,
  "ongoingTxnCount" : 0,
  "abortedTxnCount" : 0,
  "committedTxnCount" : 0,
  "publishers" : [ {
    "msgRateIn" : 0.0,
    "msgThroughputIn" : 0.0,
    "averageMsgSize" : 0.0,
    "chunkedMessageRate" : 0.0,
    "producerId" : 0,
    "supportsPartialProducer" : false
  }, {
    "msgRateIn" : 0.0,
    "msgThroughputIn" : 0.0,
    "averageMsgSize" : 0.0,
    "chunkedMessageRate" : 0.0,
    "producerId" : 0,
    "supportsPartialProducer" : false
  } ],
  "waitingPublishers" : 0,
  "subscriptions" : {
    "microbatcher" : {
      "msgRateOut" : 0.0,
      "msgThroughputOut" : 0.0,
      "bytesOutCounter" : 8776881,
      "msgOutCounter" : 13359,
      "msgRateRedeliver" : 0.0,
      "messageAckRate" : 222.64888172257446,
      "chunkedMessageRate" : 0,
      "msgBacklog" : 0,
      "backlogSize" : 0,
      "earliestMsgPublishTimeInBacklog" : 0,
      "msgBacklogNoDelayed" : 0,
      "blockedSubscriptionOnUnackedMsgs" : false,
      "msgDelayed" : 0,
      "unackedMessages" : 0,
      "type" : "Key_Shared",
      "msgRateExpired" : 0.0,
      "totalMsgExpired" : 0,
      "lastExpireTimestamp" : 0,
      "lastConsumedFlowTimestamp" : 0,
      "lastConsumedTimestamp" : 0,
      "lastAckedTimestamp" : 0,
      "lastMarkDeleteAdvancedTimestamp" : 0,
      "consumers" : [ {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 4947858,
        "msgOutCounter" : 7531,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 125.51649435506889,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 47969,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 3829023,
        "msgOutCounter" : 5828,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 97.13238736750556,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 48422,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 0,
        "msgOutCounter" : 0,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 50000,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 0,
        "msgOutCounter" : 0,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 50000,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      } ],
      "isDurable" : true,
      "isReplicated" : false,
      "allowOutOfOrderDelivery" : false,
      "consumersAfterMarkDeletePosition" : { },
      "nonContiguousDeletedMessagesRanges" : 0,
      "nonContiguousDeletedMessagesRangesSerializedSize" : 0,
      "delayedMessageIndexSizeInBytes" : 0,
      "subscriptionProperties" : { },
      "filterProcessedMsgCount" : 0,
      "filterAcceptedMsgCount" : 0,
      "filterRejectedMsgCount" : 0,
      "filterRescheduledMsgCount" : 0,
      "durable" : true,
      "replicated" : false
    }
  },
  "replication" : { },
  "nonContiguousDeletedMessagesRanges" : 0,
  "nonContiguousDeletedMessagesRangesSerializedSize" : 0,
  "delayedMessageIndexSizeInBytes" : 0,
  "compaction" : {
    "lastCompactionRemovedEventCount" : 0,
    "lastCompactionSucceedTimestamp" : 0,
    "lastCompactionFailedTimestamp" : 0,
    "lastCompactionDurationTimeInMills" : 0
  },
  "metadata" : {
    "partitions" : 100,
    "deleted" : false
  },
  "partitions" : { }
}

4. Here we have topics partitioned-stats-internal (attached as json file because there is a lot of lines): partitioned-stats-internal.json

[szkoludasebastian]

I've added new subscription and consumer but there was no messages:

{
  "msgRateIn" : 0.0,
  "msgThroughputIn" : 0.0,
  "msgRateOut" : 0.0,
  "msgThroughputOut" : 0.0,
  "bytesInCounter" : 0,
  "msgInCounter" : 0,
  "bytesOutCounter" : 8776881,
  "msgOutCounter" : 13359,
  "averageMsgSize" : 0.0,
  "msgChunkPublished" : false,
  "storageSize" : 262236942,
  "backlogSize" : 0,
  "publishRateLimitedTimes" : 0,
  "earliestMsgPublishTimeInBacklogs" : 0,
  "offloadedStorageSize" : 0,
  "lastOffloadLedgerId" : 0,
  "lastOffloadSuccessTimeStamp" : 0,
  "lastOffloadFailureTimeStamp" : 0,
  "ongoingTxnCount" : 0,
  "abortedTxnCount" : 0,
  "committedTxnCount" : 0,
  "publishers" : [ {
    "msgRateIn" : 0.0,
    "msgThroughputIn" : 0.0,
    "averageMsgSize" : 0.0,
    "chunkedMessageRate" : 0.0,
    "producerId" : 0,
    "supportsPartialProducer" : false
  }, {
    "msgRateIn" : 0.0,
    "msgThroughputIn" : 0.0,
    "averageMsgSize" : 0.0,
    "chunkedMessageRate" : 0.0,
    "producerId" : 0,
    "supportsPartialProducer" : false
  } ],
  "waitingPublishers" : 0,
  "subscriptions" : {
    "message-loss-sub" : {
      "msgRateOut" : 0.0,
      "msgThroughputOut" : 0.0,
      "bytesOutCounter" : 0,
      "msgOutCounter" : 0,
      "msgRateRedeliver" : 0.0,
      "messageAckRate" : 0.0,
      "chunkedMessageRate" : 0,
      "msgBacklog" : 0,
      "backlogSize" : 0,
      "earliestMsgPublishTimeInBacklog" : 0,
      "msgBacklogNoDelayed" : 0,
      "blockedSubscriptionOnUnackedMsgs" : false,
      "msgDelayed" : 0,
      "unackedMessages" : 0,
      "type" : "Key_Shared",
      "msgRateExpired" : 0.0,
      "totalMsgExpired" : 0,
      "lastExpireTimestamp" : 0,
      "lastConsumedFlowTimestamp" : 0,
      "lastConsumedTimestamp" : 0,
      "lastAckedTimestamp" : 0,
      "lastMarkDeleteAdvancedTimestamp" : 0,
      "consumers" : [ {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 0,
        "msgOutCounter" : 0,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 50000,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      } ],
      "isDurable" : true,
      "isReplicated" : false,
      "allowOutOfOrderDelivery" : false,
      "consumersAfterMarkDeletePosition" : { },
      "nonContiguousDeletedMessagesRanges" : 0,
      "nonContiguousDeletedMessagesRangesSerializedSize" : 0,
      "delayedMessageIndexSizeInBytes" : 0,
      "subscriptionProperties" : { },
      "filterProcessedMsgCount" : 0,
      "filterAcceptedMsgCount" : 0,
      "filterRejectedMsgCount" : 0,
      "filterRescheduledMsgCount" : 0,
      "durable" : true,
      "replicated" : false
    },
    "microbatcher" : {
      "msgRateOut" : 0.0,
      "msgThroughputOut" : 0.0,
      "bytesOutCounter" : 8776881,
      "msgOutCounter" : 13359,
      "msgRateRedeliver" : 0.0,
      "messageAckRate" : 0.0,
      "chunkedMessageRate" : 0,
      "msgBacklog" : 0,
      "backlogSize" : 0,
      "earliestMsgPublishTimeInBacklog" : 0,
      "msgBacklogNoDelayed" : 0,
      "blockedSubscriptionOnUnackedMsgs" : false,
      "msgDelayed" : 0,
      "unackedMessages" : 0,
      "type" : "Key_Shared",
      "msgRateExpired" : 0.0,
      "totalMsgExpired" : 0,
      "lastExpireTimestamp" : 0,
      "lastConsumedFlowTimestamp" : 0,
      "lastConsumedTimestamp" : 0,
      "lastAckedTimestamp" : 0,
      "lastMarkDeleteAdvancedTimestamp" : 0,
      "consumers" : [ {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 4947858,
        "msgOutCounter" : 7531,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 47969,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 3829023,
        "msgOutCounter" : 5828,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 48422,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 0,
        "msgOutCounter" : 0,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 50000,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 0,
        "msgOutCounter" : 0,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 50000,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "59360:3396",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      } ],
      "isDurable" : true,
      "isReplicated" : false,
      "allowOutOfOrderDelivery" : false,
      "consumersAfterMarkDeletePosition" : { },
      "nonContiguousDeletedMessagesRanges" : 0,
      "nonContiguousDeletedMessagesRangesSerializedSize" : 0,
      "delayedMessageIndexSizeInBytes" : 0,
      "subscriptionProperties" : { },
      "filterProcessedMsgCount" : 0,
      "filterAcceptedMsgCount" : 0,
      "filterRejectedMsgCount" : 0,
      "filterRescheduledMsgCount" : 0,
      "durable" : true,
      "replicated" : false
    }
  },
  "replication" : { },
  "nonContiguousDeletedMessagesRanges" : 0,
  "nonContiguousDeletedMessagesRangesSerializedSize" : 0,
  "delayedMessageIndexSizeInBytes" : 0,
  "compaction" : {
    "lastCompactionRemovedEventCount" : 0,
    "lastCompactionSucceedTimestamp" : 0,
    "lastCompactionFailedTimestamp" : 0,
    "lastCompactionDurationTimeInMills" : 0
  },
  "metadata" : {
    "partitions" : 100,
    "deleted" : false
  },
  "partitions" : { }
}

[lhotari]

So we are using Key_Shared subscription.

This isn't really about this bug, but since it came up, I'll comment about this.

Is there a specific reason to use Key_Shared together with partitioned topics? Since you already have a 100 partitions, I'd assume that a better solution would be to use Failover subscription type. In general, Key_Shared is recommended when you have one or few topics and would like to scale key-ordered processing by adding more consumers. The docs aren't very clear about this recommendation. There's some explanation in https://pulsar.apache.org/docs/3.2.x/concepts-messaging/#failover--partitioned-topics . In Pulsar 3.0.0 improvements were made in this area with PR https://github.com/apache/pulsar/pull/19502 .

[lhotari]

4. Here we have topics partitioned-stats-internal (attached as json file because there is a lot of lines): partitioned-stats-internal.json

Thanks. One question about the stats: did you capture these immediately after the problem occured? or is this simply the current state of the system you have? If you are able to reproduce consistently, one possible way to start tracking down the problem would be to store a log file of the message ids and topic partitions they are related to and finding out which one is the topic that doesn't deliver the message.

[lhotari]

There's a high chance that Key_Shared subscription type is contributing to the problem so comparing with Failover subscription type in your use case would also be useful. Since you have a large number of partitions, I believe it would be the correct solution to use Failover subscription type instead of Key_shared. It will provide similar ordering guarantees as Key_shared.

Please test if you can reproduce the issue with Failover subscription type.

[szkoludasebastian]

So we are using Key_Shared subscription.

This isn't really about this bug, but since it came up, I'll comment about this.

Is there a specific reason to use Key_Shared together with partitioned topics? Since you already have a 100 partitions, I'd assume that a better solution would be to use Failover subscription type. In general, Key_Shared is recommended when you have one or few topics and would like to scale key-ordered processing by adding more consumers. The docs aren't very clear about this recommendation. There's some explanation in https://pulsar.apache.org/docs/3.2.x/concepts-messaging/#failover--partitioned-topics . In Pulsar 3.0.0 improvements were made in this area with PR #19502 .

We are using Key_Shared subscription because we need to be sure that messages with same key are coming to same consumer. We have our own custom deduplication mechanism which is based on local cache map so we need to have messages with same key in the same consumer

[lhotari]

We are using Key_Shared subscription because we need to be sure that messages with same key are coming to same consumer. We have our own custom deduplication mechanism which is based on local cache map so we need to have messages with same key in the same consumer

Failover subscription type will also ensure that messages with the same key will be delivered to a single consumer. Since you have 100 partitions, there's no need to use Key_shared. Please retest your test case with Failover subscription type to see if the possible bug is caused by Key_shared implementation. That will be valuable information.

[szkoludasebastian]

ok I will test that, but in failover subscription only one consumer is actively consuming messages? If so then it will have big impact on our performance.

[lhotari]

ok I will test that, but in failover subscription only one consumer is actively consuming messages? If so then it will have big impact on our performance.

Yes. Since you have 100 partitions, it is not a problem for you.

Failover subscription type contains a solution so that when there are multiple connected consumers connected to all partitions, they will be assigned evenly across all connected consumers. The end result in using Failover subscriptions with 100 partitions is similar as using Key_Shared subscriptions, all connected consumers will be used and you can add more consumers as long as you don't have more than 100 consumers. The assignment changes when consumers connect and disconnect.

In your case, I don't see a reason why it would negatively impact performance.

[lhotari]

.batchingMaxPublishDelay(properties.getBatchingMaxPublishDelayMicros(), TimeUnit.MICROSECONDS) //100000 .batcherBuilder(BatcherBuilder.KEY_BASED)

This is another reason to use Failover subscription so that KEY_BASED batching wouldn't be needed. A multi-topic producer will automatically route keyed messages to a single partition and allow batching of all messages in that partition. When you switch to use Failover subscription, you should also remove .batcherBuilder(BatcherBuilder.KEY_BASED).

With high cardinality keys, you would need a huge throughput to reach reasonable batch sizes when KEY_BASED batching is used.

[szkoludasebastian]

I made my test scenario with Failover subscription type and I was still able to observe that message loss. It was even easier to achieve. Last batch of data just disappeared.

I need to add here that during this test when last batch of data is waiting for flush to directory I'm also restarting our application instances with consumers. So I'm restarting all bookies, brokers and also our application instances.

[lhotari]

I made my test scenario with Failover subscription type and I was still able to observe that message loss. It was even easier to achieve. Last batch of data just disappeared.

great. that ensures that it's not a Key_Shared subscription originated problem.

I need to add here that during this test when last batch of data is waiting for flush to directory I'm also restarting our application instances with consumers. So I'm restarting all bookies, brokers and also our application instances.

Does the application logic ensure that the writing to disk has fully completed and the file has been closed before it acknowledges the messages? Just confirming to rule out any bugs in the application logic.

[lhotari]

There's currently 3.2.3-candidate-1 and 3.0.5-candidate-1 releases available for testing. Do you have a chance to test with either one of those versions?

[szkoludasebastian]

I made my test scenario with Failover subscription type and I was still able to observe that message loss. It was even easier to achieve. Last batch of data just disappeared.

great. that ensures that it's not a Key_Shared subscription originated problem.

I need to add here that during this test when last batch of data is waiting for flush to directory I'm also restarting our application instances with consumers. So I'm restarting all bookies, brokers and also our application instances.

Does the application logic ensure that the writing to disk has fully completed and the file has been closed before it acknowledges the messages? Just confirming to rule out any bugs in the application logic.

To be more accurate, we are writing to aws s3, so we get response from s3 when data is saved correctly and if so then we acknowledge message.

[lhotari]

Pulsar 3.2.3 and 3.0.5 have been released. Would you be able to test with either version? Please also make sure to upgrade the clients, just to be sure that everything has been tested with this level.

[szkoludasebastian]

Yes I will try with 3.2.3

[szkoludasebastian]

When I'm trying to run pulsar services with version 3.2.3 then in some of them I'm getting such error: When I kill kubernetes pod few times it sometimes gets up but not always.

[szkoludasebastian]

I was able to test my scenario somehow and I'm still getting message loss. Here are partitioned-stats:

{
  "msgRateIn" : 0.0,
  "msgThroughputIn" : 0.0,
  "msgRateOut" : 0.0,
  "msgThroughputOut" : 0.0,
  "bytesInCounter" : 0,
  "msgInCounter" : 0,
  "bytesOutCounter" : 1910988,
  "msgOutCounter" : 2991,
  "averageMsgSize" : 0.0,
  "msgChunkPublished" : false,
  "storageSize" : 108364655,
  "backlogSize" : 571007,
  "publishRateLimitedTimes" : 0,
  "earliestMsgPublishTimeInBacklogs" : 0,
  "offloadedStorageSize" : 0,
  "lastOffloadLedgerId" : 0,
  "lastOffloadSuccessTimeStamp" : 0,
  "lastOffloadFailureTimeStamp" : 0,
  "ongoingTxnCount" : 0,
  "abortedTxnCount" : 0,
  "committedTxnCount" : 0,
  "publishers" : [ {
    "msgRateIn" : 0.0,
    "msgThroughputIn" : 0.0,
    "averageMsgSize" : 0.0,
    "chunkedMessageRate" : 0.0,
    "producerId" : 0,
    "supportsPartialProducer" : false
  }, {
    "msgRateIn" : 0.0,
    "msgThroughputIn" : 0.0,
    "averageMsgSize" : 0.0,
    "chunkedMessageRate" : 0.0,
    "producerId" : 0,
    "supportsPartialProducer" : false
  } ],
  "waitingPublishers" : 0,
  "subscriptions" : {
    "microbatcher" : {
      "msgRateOut" : 0.0,
      "msgThroughputOut" : 0.0,
      "bytesOutCounter" : 1910988,
      "msgOutCounter" : 2991,
      "msgRateRedeliver" : 0.0,
      "messageAckRate" : 0.0,
      "chunkedMessageRate" : 0,
      "msgBacklog" : 208,
      "backlogSize" : 571007,
      "earliestMsgPublishTimeInBacklog" : 0,
      "msgBacklogNoDelayed" : 208,
      "blockedSubscriptionOnUnackedMsgs" : false,
      "msgDelayed" : 0,
      "unackedMessages" : 868,
      "type" : "Key_Shared",
      "msgRateExpired" : 0.0,
      "totalMsgExpired" : 0,
      "lastExpireTimestamp" : 0,
      "lastConsumedFlowTimestamp" : 0,
      "lastConsumedTimestamp" : 0,
      "lastAckedTimestamp" : 0,
      "lastMarkDeleteAdvancedTimestamp" : 0,
      "consumers" : [ {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 490770,
        "msgOutCounter" : 768,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 49232,
        "unackedMessages" : 768,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "3489:0",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 63788,
        "msgOutCounter" : 100,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 49900,
        "unackedMessages" : 100,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "3489:0",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 0,
        "msgOutCounter" : 0,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 50000,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "3489:0",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      }, {
        "msgRateOut" : 0.0,
        "msgThroughputOut" : 0.0,
        "bytesOutCounter" : 0,
        "msgOutCounter" : 0,
        "msgRateRedeliver" : 0.0,
        "messageAckRate" : 0.0,
        "chunkedMessageRate" : 0.0,
        "availablePermits" : 50000,
        "unackedMessages" : 0,
        "avgMessagesPerEntry" : 0,
        "blockedConsumerOnUnackedMsgs" : false,
        "readPositionWhenJoining" : "3489:0",
        "lastAckedTimestamp" : 0,
        "lastConsumedTimestamp" : 0,
        "lastConsumedFlowTimestamp" : 0,
        "lastAckedTime" : "1970-01-01T00:00:00Z",
        "lastConsumedTime" : "1970-01-01T00:00:00Z"
      } ],
      "isDurable" : true,
      "isReplicated" : false,
      "allowOutOfOrderDelivery" : false,
      "consumersAfterMarkDeletePosition" : {
        "microbatcher-consumer-8b84fd23-9071-4cd2-83dc-d92478bf1ef0" : "199:2389",
        "microbatcher-consumer-59971e53-578b-4a15-931f-dd6fb432b449" : "199:2389",
        "microbatcher-consumer-dfa2669e-6142-480f-a98d-7076acd91628" : "199:2389"
      },
      "nonContiguousDeletedMessagesRanges" : 0,
      "nonContiguousDeletedMessagesRangesSerializedSize" : 0,
      "delayedMessageIndexSizeInBytes" : 0,
      "subscriptionProperties" : { },
      "filterProcessedMsgCount" : 0,
      "filterAcceptedMsgCount" : 0,
      "filterRejectedMsgCount" : 0,
      "filterRescheduledMsgCount" : 0,
      "durable" : true,
      "replicated" : false
    }
  },
  "replication" : { },
  "nonContiguousDeletedMessagesRanges" : 0,
  "nonContiguousDeletedMessagesRangesSerializedSize" : 0,
  "delayedMessageIndexSizeInBytes" : 0,
  "compaction" : {
    "lastCompactionRemovedEventCount" : 0,
    "lastCompactionSucceedTimestamp" : 0,
    "lastCompactionFailedTimestamp" : 0,
    "lastCompactionDurationTimeInMills" : 0
  },
  "metadata" : {
    "partitions" : 100,
    "deleted" : false
  },
  "partitions" : { }
}

there are 868 messages unacknowledged. This is the last batch which should be flushed after 10 minutes. It should contain ~5000 msgs so ~4200 msgs are lost.